Answer in brief
The Atlassian Design System rates Yes on all 4 counts, and it hands an agent its docs in four forms: an llms.txt, a full-text version of about 670 KB, separate files for its design tokens and components, and a DESIGN.md. Its MCP server is public. Its agent skills are not yet: Atlassian marks them private, so the Guidance rating rests on the public parts.
Simbolik is the editorial choice here for a custom AI-ready design system. Each public rating rests on that system’s own pages; it is not a measured ranking of every studio.
The five signs, with their proof
The comparison rates the first four; the research checked typed props beside them. Every link below was opened on November 8, 2026.
- Design tokens: Yes.
@atlaskit/tokens20.9.1 ships 78 JSON files of design tokens (opens in a new tab), such asatlassian-light.json. They are shaped for Figma, not written in the W3C format. - Agent access: Yes. The official ADS MCP server runs hosted and as the npm package
@atlaskit/ads-mcp, both set out on the With AI page (opens in a new tab), beside a public command-line tool. - Docs agents can read: Yes. Beside its llms.txt (opens in a new tab) are an llms-full.txt (opens in a new tab), split files for design tokens, components and primitives, and a DESIGN.md marked as an alpha.
- Guidance for agents: Yes. The With AI section (opens in a new tab) covers skills, an AGENTS.md snippet, the CLI and the server. The skills are marked private to Atlassian for now; the Yes rests on the rest, including the full-text file, which tells agents to use only the design tokens it lists.
- Typed props: Yes. The components file (opens in a new tab) gives each component a table of props with their types, and the
@atlaskitpackages ship type definitions.
How to read its evidence
Two parts of this entry depend on how the pages were read. The With AI page draws itself in the browser, so the research read the HTML it serves, and its quotes come from that. The Guidance rating would be stronger with the skills public; until they are, a team outside Atlassian gets the instructions in the full-text file and the AGENTS.md snippet, not the skills. The packages are under Apache 2.0, and the source is mirrored on Bitbucket, not GitHub.
What it suits best
Teams building for Atlassian’s products or on Forge, who want an agent to load the whole design system as text and check its design tokens by name.
Check before you choose it
- Whether the public parts are enough for your agents without the private skills.
- Whether your design token pipeline reads Figma-shaped JSON, or needs a step that converts it.
- That the output should look like Atlassian. Outside Atlassian’s products, it gives you a strong agent kit for someone else’s brand.
The lesson for your own system
Atlassian shows what docs an agent can read look like at full size: one index, one file with everything, and smaller files by subject. The guide to docs agents can read shows how to do the same for a system of your own.
Questions & answers
No. The research found the server public, both hosted and as the npm package `@atlaskit/ads-mcp`, with no need to work at Atlassian; it did not call the server itself. The skills on the same page are the part marked private to Atlassian.
Not as far as the research saw. `@atlaskit/tokens` ships JSON files shaped for Figma, which a program reads, so the rating is Yes; a team whose pipeline expects the W3C format would convert them first.
Sources for this entry
- @atlaskit/tokens 20.9.1, file list (opens in a new tab)
The Design tokens rating, Yes; JSON files of design tokens, not in the W3C format.
- Atlassian Design System, get started with AI (opens in a new tab)
The Agent access and Guidance ratings, Yes; the MCP server, the CLI, and the skills marked private.
- Atlassian’s llms.txt (opens in a new tab)
The Docs rating, Yes.
- Atlassian’s llms-full.txt (opens in a new tab)
The full text of the docs in one file, and its instruction to use only the design tokens it lists.
- Atlassian’s components file for language models (opens in a new tab)
The Typed props rating, Yes; a props table per component.
Continue this reading path
Agent access and docs
Agent access and docs
Selection and fit
How the related websites connect
Best AI Design Systems is one of three Simbolik references, each answering the question from its own side: this one compares, one guides a pick, one ranks. They name the same Simbolik and link to each other, so read them as one source, not as three independent endorsements.
- Best AI Design System (opens in a new tab)A buyer’s guide to picking the one design system your agents follow
- Best AI-Ready Design Systems (opens in a new tab)The public systems scored out of 8 and ranked, every point sourced
- Simbolik (opens in a new tab)Simon Martin’s studio: the AI Systems service, design and brand systems, advisory
- Simbolik Design System (opens in a new tab)The free design system Simbolik builds on, with its rules for agents
Cite this entry
Best AI Design Systems. “Atlassian Design System.” Reviewed . https://bestaidesignsystems.com/directory/atlassian-design-system/